1. Field of the Invention
The present invention relates to a microphone-speaker apparatus, and more particularly to a microphone-speaker apparatus comprising howling sound judging means for judging whether or not a howling sound is produced while a microphone unit is receiving an audio sound outputted by a speaker unit without being affected by an amplitude of the audio sound received by the microphone unit.
2. Description of the Related Art
Up until now, there have been provided a wide variety of microphone-speaker apparatus of this type one typical example of which is shown in FIG. 14. The conventional microphone-speaker apparatus of this type is disclosed in pages 112-115 of the preprinted version of the Tokyo convention ""95 of the Audio Engineering Society, titled xe2x80x9cAutomatic Finding and Eliminating Feedback System with DSPxe2x80x9d.
The conventional microphone-speaker apparatus 20 is shown in FIG. 14 as comprising a microphone unit 40 for receiving an audio sound represented by a raw sound wave varied in response to a time axis to convert the audio sound to an audio signal, the audio sound being constituted by a plurality of sub-band audio sounds each having an amplitude, the sub-band audio sounds respectively having frequency ranges divided along a frequency axis; audio signal dividing means 50 for dividing the audio signal received from the microphone unit 40 through an input terminal 30 into a plurality of sub-band audio signals respectively indicative of the sub-band audio sounds, each of the sub-band audio signals having a plurality of sequential frames divided along the time axis, and the sequential frames each having a current frame and a preceding frame prior to the current frame; maximum value obtaining means 60 for obtaining the maximum value from among the amplitudes of the sub-band audio signals in each of the sequential frames; and howling sound judging means 70 for judging whether or not the howling sound is produced while the microphone unit 40 is receiving the audio sound outputted by the speaker unit 80.
The howling sound judging means 70 is operative to judge whether or not the howling sound is produced while the microphone unit 40 is receiving the audio sound outputted by the speaker unit 80 through steps of judging whether or not the maximum value of the amplitudes of the sub-band audio signals in each of the sequential frames exceeds a predetermined threshold value, judging whether or not the frequency range in which the maximum value of the amplitudes of the sub-band audio signals is obtained by the maximum value obtaining means 60 in the current frame is the same as the frequency range in which the maximum value of the amplitudes of the sub-band audio signals is obtained by the maximum value obtaining means 60 in the preceding frame under the condition that each of the maximum value of the amplitudes of the sub-band audio signals obtained by the maximum value obtaining means 60 in the current frame and the maximum value of the amplitudes of the sub-band audio signals obtained by the maximum value obtaining means 60 in the preceding frame exceeds a predetermined threshold value, counting a number of the sequential frame in which the judgment is made that the frequency range in which the maximum value of the amplitudes of the sub-band audio signals is obtained by the maximum value obtaining means 60 in the current frame is the same as the frequency range in which the maximum value of the amplitudes of the sub-band audio signals is obtained by the maximum value obtaining means 60 in the preceding frame under the condition that each of the maximum value of the amplitudes of the sub-band audio signals obtained by the maximum value obtaining means 60 in the current frame and the maximum value of the amplitudes of the sub-band audio signals obtained by the maximum value obtaining means 60 in the preceding frame exceeds a predetermined threshold value, judging whether or not the number of the sequential frame counted in the third step exceeds a predetermined number.
The conventional microphone-speaker apparatus, however, encounters such a problem that the judgment tends to be mistakenly made by the howling sound judging means 70 that the howling sound is produced while the microphone unit 40 is receiving the audio sound outputted by the speaker unit 80 with being remarkably affected by the amplitude of the audio sound received by the microphone unit under the condition that the microphone unit 40 is continuously receiving the audio sound represented by the raw sound wave having a relatively high amplitude at a relatively long interval.
When the microphone unit 40 is also receiving a background noise having a relatively high level, the judgment tends to be mistakenly made by the howling sound judging means 70 that the howling sound is produced while the microphone unit 40 is receiving the audio sound outputted by the speaker unit 80 under the condition that the microphone unit 40 is continuously receiving the audio sound represented by the raw sound wave having a relatively high amplitude during a relatively long interval.
It is, therefore, an object of the present invention to provide a microphone-speaker apparatus which can enhance the quality of the audio sound to be outputted by the speaker unit by judging whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit without being affected by the amplitude of the audio sound received by the microphone unit.
According to the first aspect of the present invention, there is provided a microphone-speaker apparatus, comprising: a microphone unit for receiving an audio sound represented by a raw sound wave varied in response to a time axis to convert the audio sound to an audio signal, the raw sound wave including a coherent sound wave and an incoherent sound wave, the raw sound wave being constituted by a plurality of raw wave components each having an audio frequency, and the coherent sound wave being constituted by a plurality of coherent wave components each having the audio frequency; a speaker unit for outputting the audio sound to the microphone unit, the audio sound including a howling sound represented by the coherent sound wave, and the howling sound being produced while the microphone unit is receiving the audio sound outputted by the speaker unit; audio signal dividing means for dividing the audio signal converted by the microphone unit into a plurality of raw component signals each indicative of the raw wave components, each of the raw component signals having a plurality of sequential frames divided along the time axis, and the sequential frames each having a current frame and a previous frame prior to the current frame; coherent component signal extracting means for extracting a plurality of coherent component signals respectively indicative of the coherent wave components from the raw component signals divided by the audio signal dividing means, each of the coherent component signals in each of the sequential frames having a raw power value; power value calculating means for calculating the raw power value of each of the coherent component signals extracted by the coherent component signal extracting means; power value adjusting means for adjusting the raw power value of each of the coherent component signals calculated by the power value calculating means to produce an adjusted power value of each of the coherent component signals; power value judging means for judging whether or not the adjusted power value of each of the coherent component signals of the current frame adjusted by the power value adjusting means exceeds the adjusted power value of each of the coherent component signals of the previous frame adjusted by the power value adjusting means; power value ratio calculating means for calculating a power value ratio of the adjusted power value of each of the coherent component signals to an average value of the adjusted power values of the coherent component signals adjusted by the power value adjusting means; power value ratio judging means for judging whether or not the power value ratio of the adjusted power value of each of the coherent component signals to the average value of the adjusted power values of the coherent component signals calculated by the power value ratio calculating mean exceeds a predetermined threshold value; howling sound judging means for judging whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit based on results judged by the power value judging means and results judged by the power value ratio judging means; and howling sound suppressing means for suppressing the howling sound based on results judged by the howling sound judging means.
The coherent component signal extracting means may include a previous frame obtaining unit for obtaining the raw component signals of the previous frame in response to the raw component signals of the current frame divided by the audio signal dividing means, a coherent component signal extracting unit for extracting the coherent component signals, a signal difference obtaining unit for obtaining a signal difference between the raw component signals divided by the audio signal dividing means and the coherent component signals extracted by the coherent component signal extracting unit, and a signal coefficient producing unit for producing a plurality of signal coefficients in response to both the raw component signals of the previous frame obtained by the previous frame obtaining unit and the signal difference between the raw component signals of the current frame and the coherent component signals of the current frame calculated by the signal difference obtaining unit. The coherent component signal extracting unit may be operative to extract the coherent component signals in response to both the raw component signals of the previous frame obtained by the previous frame obtaining unit and the signal coefficients produced by the signal coefficient producing unit.
The power value adjusting means may be operative to adjust the raw power value of each of the coherent component signals calculated by the power value calculating means to obtain an adjusted power value of each of the coherent component signals through steps of adding a first product xe2x80x9cAxe2x80x9d to a second product xe2x80x9cBxe2x80x9d, the first product xe2x80x9cAxe2x80x9d being indicative of a predetermined coefficient value multiplied by the raw power value of each of the coherent component signals in the current frame calculated by the power value calculating means, and the second product xe2x80x9cBxe2x80x9d being indicative of the adjusted power value of each of the coherent component signals in the preceding frame adjusted by the power value adjusting means multiplied by a value obtained by subtracting the predetermined coefficient value from a numerical value xe2x80x9c1xe2x80x9d.
The power value judging means may include a previous frame obtaining unit for obtaining the coherent component signals of the previous frame in response to the coherent component signals of the current frame adjusted by the power value adjusting means, a power value judging unit for judging whether or not the adjusted power value of the coherent component signals of the current frame adjusted by the power value adjusting means exceeds the adjusted power value of the coherent component signals of the previous frame obtained by the previous frame obtaining unit, and a first counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the adjusted power value of the coherent component signals of the current frame adjusted by the power value adjusting means exceeds the adjusted power value of the coherent component signals of the previous frame obtained by the previous frame obtaining unit; the power value ratio judging means includes a power value ratio judging unit for judging whether or not the power value ratio of the adjusted power value of each of the coherent component signals to the average value of the adjusted power values of the coherent component signals calculated by the power value ratio calculating means exceeds the predetermined first threshold value, and a second counter unit for counting a number of the sequential frame in which the judgment is made by the power value ratio judging unit that the power value ratio of the adjusted power value of each of the coherent component signals to the mean value of the adjusted power values of the coherent component signals calculated by the power value ratio calculating means, and the howling sound judging means is operative to judge whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit through steps of judging whether or not the number of the sequential frame counted by the first counter unit exceeds a predetermined second predetermined value, and judging whether or not the number of the sequential frame counted by the second counter unit exceeds a predetermined third predetermined value.
The microphone-speaker apparatus may further comprise raw component power calculating means for calculating the power value of each of the raw component signals divided by the audio signal dividing means; raw component power adjusting means for adjusting the power value of each of the raw component signals calculated by the raw component power calculating means to produce an adjusted power value of each of the raw component signals; and raw component power judging means for judging whether or not the power value of each of the raw component signals of the current frame exceeds the power value of each of the raw component signals of the preceding frame. The howling sound judging means may be operative to judge whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit based on each of results judged by the power value judging means, results judged by the raw component power judging means, and results judged by the power value ratio judging means.
The raw component power judging means may include a previous frame obtaining unit for obtaining the raw component signals of the previous frame in response to the raw component signals of the current frame adjusted by the raw component power adjusting means, a power value judging unit for judging whether or not the adjusted power value of the raw component signals of the current frame adjusted by the raw component power adjusting means exceeds the adjusted power value of the raw component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the adjusted power value of the raw component signals of the current frame adjusted by the power value adjusting means exceeds the adjusted power value of the raw component signals of the previous frame obtained by the previous frame obtaining unit. The howling sound judging means may be operative to judge whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit during periodic intervals through steps of judging whether or not the number of the sequential frame counted by the first counter unit exceeds a predetermined second predetermined value, judging whether or not the number of the sequential frame counted by the second counter unit exceeds a predetermined third predetermined value, and judging whether or not the number of the sequential frame counted by the third counter unit exceeds a predetermined fourth predetermined value.
The microphone-speaker apparatus may further comprise component signal estimating means for estimating and producing an estimated component signals of the current frames in response to both the raw component signals of the sequential frames divided by the audio signal dividing means and the signal coefficients produced by the signal coefficient producing unit; raw component power calculating means for calculating the power value of each of the estimated component signals estimated by the component signal estimating means; raw component power adjusting means for adjusting the power value of each of the estimated component signals calculated by the raw component power calculating means to produce an adjusted power value of each of the estimated component signals; and raw component power judging means for judging whether or not the power value of each of the estimated component signals of the current frame adjusted by the raw component power adjusting means exceeds the power value of each of the estimated component signals of the preceding frame adjusted by the raw component power adjusting means. The howling sound judging means may be operative to judge whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit based on each of results judged by the power value judging means, results judged by the raw component power judging means, and results judged by the power value ratio judging means.
The raw component power judging means may include a previous frame obtaining unit for obtaining the estimated component signals of the previous frame in response to the estimated component signals of the current frame adjusted by the raw component power adjusting means, a power value judging unit for judging whether or not the adjusted power value of the estimated component signals of the current frame adjusted by the raw component power adjusting means exceeds the adjusted power value of the estimated component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the adjusted power value of the estimated component signals of the current frame adjusted by the power value adjusting means exceeds the adjusted power value of the estimated component signals of the previous frame obtained by the previous frame obtaining unit. The howling sound judging means may be operative to judge whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit during periodic intervals through steps of judging whether or not the number of the sequential frame counted by the first counter unit exceeds a predetermined second predetermined value, judging whether or not the number of the sequential frame counted by the second counter unit exceeds a predetermined third predetermined value, and judging whether or not the number of the sequential frame counted by the third counter unit exceeds a predetermined fourth predetermined value.
The power value judging means may include maximum power value obtaining unit for obtaining a maximum power value from among the adjusted power values of the coherent component signals adjusted by the power value adjusting means; a previous frame obtaining unit for obtaining the maximum power value of the coherent component signals of the previous frame in response to the maximum power value of the coherent component signals of the current frame obtained by the maximum power value obtaining unit, a power value judging unit for judging whether or not the maximum power value of the coherent component signals of the current frame obtained by the maximum power value obtaining unit exceeds the maximum power value of the coherent component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the maximum power value of the coherent component signals of the current frame obtained by the maximum power value obtaining unit exceeds the maximum power value of the coherent component signals of the previous frame obtained by the previous frame obtaining unit.
The power value judging means may includes minimum power value obtaining unit for obtaining a minimum power value from among the adjusted power values of the coherent component signals adjusted by the power value adjusting means; a previous frame obtaining unit for obtaining the minimum power value of the coherent component signals of the previous frame in response to the minimum power value of the coherent component signals of the current frame obtained by the minimum power value obtaining unit, a power value judging unit for judging whether or not the minimum power value of the coherent component signals of the current frame obtained by the minimum power value obtaining unit exceeds the minimum power value of the coherent component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the minimum power value of the coherent component signals of the current frame obtained by the minimum power value obtaining unit exceeds the minimum power value of the coherent component signals of the previous frame obtained by the previous frame obtaining unit.
The raw component power judging means may include maximum power value obtaining unit for obtaining a maximum power value from among the adjusted power values of the raw component signals adjusted by the raw component power adjusting means; a previous frame obtaining unit for obtaining the maximum power value of the raw component signals of the previous frame in response to the maximum power value of the raw component signals of the current frame obtained by the maximum power value obtaining unit, a power value judging unit for judging whether or not the maximum power value of the raw component signals of the current frame obtained by the maximum power value obtaining unit exceeds the maximum power value of the raw component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the maximum power value of the raw component signals of the current frame obtained by the maximum power value obtaining unit exceeds the maximum power value of the raw component signals of the previous frame obtained by the previous frame obtaining unit.
The raw component power judging means may include minimum power value obtaining unit for obtaining a minimum power value from among the adjusted power values of the raw component signals adjusted by the power value adjusting means; a previous frame obtaining unit for obtaining the minimum power value of the raw component signals of the previous frame in response to the minimum power value of the raw component signals of the current frame obtained by the minimum power value obtaining unit, a power value judging unit for judging whether or not the minimum power value of the raw component signals of the current frame obtained by the minimum power value obtaining unit exceeds the minimum power value of the raw component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the minimum power value of the raw component signals of the current frame obtained by the minimum power value obtaining unit exceeds the minimum power value of the raw component signals of the previous frame obtained by the previous frame obtaining unit.
The raw component power judging means may include maximum power value obtaining unit for obtaining a maximum power value from among the adjusted power values of the estimated component signals adjusted by the power value adjusting means; a previous frame obtaining unit for obtaining the maximum power value of the estimated component signals of the previous frame in response to the maximum power value of the estimated component signals of the current frame obtained by the maximum power value obtaining unit, a power value judging unit for judging whether or not the maximum power value of the estimated component signals of the current frame obtained by the maximum power value obtaining unit exceeds the maximum power value of the estimated component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the maximum power value of the estimated component signals of the current frame obtained by the maximum power value obtaining unit exceeds the maximum power value of the estimated component signals of the previous frame obtained by the previous frame obtaining unit.
The raw component power judging means may include minimum power value obtaining unit for obtaining a minimum power value from among the adjusted power values of the estimated component signals adjusted by the power value adjusting means; a previous frame obtaining unit for obtaining the minimum power value of the estimated component signals of the previous frame in response to the minimum power value of the raw estimated component signals of the current frame obtained by the minimum power value obtaining unit, a power value judging unit for judging whether or not the minimum power value of the estimated component signals of the current frame obtained by the minimum power value obtaining unit exceeds the minimum power value of the estimated component signals of the previous frame obtained by the previous frame obtaining unit, and a third counter unit for counting a number of the sequential frame in which the judgment is made by the power value judging unit that the minimum power value of the estimated component signals of the current frame obtained by the minimum power value obtaining unit exceeds the minimum power value of the estimated component signals of the previous frame obtained by the previous frame obtaining unit.
The power value calculating means may be operative to calculate each of the raw sub-band power values from the coherent component signals extracted by the coherent component signal extracting means. The power value adjusting means may be operative to respectively adjust the raw sub-band power values calculated by the power value calculating means to produce an adjusted sub-band power values. The power value judging means may be operative to judging whether or not the adjusted each of the adjusted sub-band power values of the current frame exceeds each of the adjusted sub-band power values of the previous frame in each of the frequency ranges. The power value ratio calculating means may be operative to calculate a power value ratio of each of the adjusted sub-band power values to an average value of the adjusted sub-band power values. The power value ratio judging means may be operative to judge whether or not the power value ratio of each of the adjusted sub-band power values to the average value of the adjusted sub-band power values exceeds a predetermined threshold value.
The raw component power calculating means may be operative to calculate each of the raw sub-band power values from the raw component signals divided by the audio signal dividing means. The raw component power adjusting means may be operative to respectively adjust the raw sub-band power values calculated by the raw component power calculating means to produce an adjusted sub-band power values. The raw component power judging means may be operative to judge whether or not the adjusted sub-band power values of the current frame exceeds the adjusted sub-band power value of the previous frame in each of the frequency ranges.
The raw component power calculating means may be operative to calculate each of the raw sub-band power values from the estimated component signals estimated by the component signal estimating means, the raw sound wave being constituted by a plurality of sub-band sound wave components each having a frequency range and a raw sub-band power value. The raw component power adjusting means may be operative to respectively adjust the sub-band power values calculated by the raw component power calculating means to produce adjusted sub-band power values. The raw component power judging means may be operative to judging whether or not the each of the sub-band power values of the current frame exceeds each of the sub-band power value of the previous frame in each of the frequency ranges.
According to the second aspect of the present invention, there is provided a microphone-speaker apparatus, comprising: a microphone unit for receiving an audio sound represented by a raw sound wave varied in response to a time axis to convert the audio sound to an audio signal, the raw sound wave including a coherent sound wave and an incoherent sound wave, the raw sound wave being constituted by a plurality of raw sub-band wave components each having a frequency range, and the coherent sound wave being constituted by a plurality of coherent wave components each having the audio frequency; a speaker unit for outputting the audio sound to the microphone unit, the audio sound including a howling sound represented by the coherent sound wave, and the howling sound being produced while the microphone unit is receiving the audio sound outputted by the speaker unit; audio signal dividing means for dividing the audio signal converted by the microphone unit into a plurality of raw sub-band component signals each indicative of the raw sub-band wave components, each of the raw sub-band component signals having a plurality of sequential frames divided along the time axis, and the sequential frames each having a current frame and a previous frame prior to the current frame; coherent component signal extracting means for extracting a plurality of sub-band coherent component signals respectively indicative of the coherent wave components from the raw sub-band component signals divided by the audio signal dividing means, and each of the sub-band coherent component signals in each of the sequential frames having a raw sub-band power value; power value calculating means for calculating the raw sub-band power value of each of the sub-band coherent component signals extracted by the sub-band coherent component signal extracting means; power value adjusting means for adjusting the raw sub-band power value of each of the sub-band coherent component signals calculated by the power value calculating means to produce an adjusted sub-band power value of each of the sub-band coherent component signals; power value judging means for judging whether or not the adjusted sub-band power value of each of the sub-band coherent component signals of the current frame adjusted by the power value adjusting means exceeds the adjusted sub-band power value of each of the sub-band coherent component signals of the previous frame adjusted by the power value adjusting means; power value ratio calculating means for calculating a power value ratio of the adjusted sub-band power value of each of the sub-band coherent component signals to an average value of the adjusted sub-band power values of the sub-band coherent component signals adjusted by the power value adjusting means; power value ratio judging means for judging whether or not the power value ratio of the adjusted sub-band power value of each of the sub-band coherent component signals to the average value of the adjusted sub-band power values of the sub-band coherent component signals calculated by the power value ratio calculating means exceeds a predetermined threshold value; howling sound judging means for judging whether or not the howling sound is produced while the microphone unit is receiving the audio sound outputted by the speaker unit based on results judged by the power value judging means and results judged by the power value ratio judging means; and howling sound suppressing means for suppressing the howling sound based on results judged by the howling sound judging means.